view contrib/relnotes @ 48178:f12a19d03d2c

fix: reduce number of tool executions

By grouping together (path, ctx) pairs according to the inputs they would
provide to fixer tools, we can deduplicate executions of fixer tools to
significantly reduce the amount of time spent running slow tools.

This change does not handle clean files in the working copy, which could still
be deduplicated against the files in the checked out commit. It's a little
harder to do that because the filerev is not available in the workingfilectx
(and it doesn't exist for added files).

Anecdotally, this change makes some real uses cases at Google 10x faster. I
think we were originally hesitant to do this because the benefits weren't
obvious, and implementing it efficiently is kind of tricky. If we simply
memoized the formatter execution function, we would be keeping tons of file
content in memory.

Also included is a regression test for a corner case that I broke with my first
attempt at optimizing this code.

Differential Revision: https://phab.mercurial-scm.org/D11280

#!/usr/bin/env python3
"""Generate release notes from our commit log.

This uses the relnotes extension directives when they're available,
and falls back to our old pre-relnotes logic that used to live in the
release-tools repo.
"""
import argparse
import re
import subprocess

rules = {
    # keep
    r"\(issue": 100,
    r"\(BC\)": 100,
    r"\(API\)": 100,
    r"\(SEC\)": 100,
    # core commands, bump up
    r"(commit|files|log|pull|push|patch|status|tag|summary)(|s|es):": 20,
    r"(annotate|alias|branch|bookmark|clone|graft|import|verify).*:": 20,
    # extensions, bump up
    r"(mq|shelve|rebase):": 20,
    # newsy
    r": deprecate": 20,
    r": new.*(extension|flag|module)": 10,
    r"( ability|command|feature|option|support)": 10,
    # experimental
    r"hg-experimental": 20,
    r"(from|graduate).*experimental": 15,
    r"(hide|mark).*experimental": -10,
    # bug-like?
    r"(fix|don't break|improve)": 7,
    r"(not|n't|avoid|fix|prevent).*crash": 10,
    r"vulnerab": 10,
    # boring stuff, bump down
    r"^contrib": -5,
    r"debug": -5,
    r"help": -5,
    r"minor": -5,
    r"(doc|metavar|bundle2|obsolete|obsmarker|rpm|setup|debug\S+:)": -15,
    r"(check-code|check-commit|check-config|import-checker)": -20,
    r"(flake8|lintian|pyflakes|pylint)": -20,
    # cleanups and refactoring
    r"(clean ?up|white ?space|spelling|quoting)": -20,
    r"(flatten|dedent|indent|nesting|unnest)": -20,
    r"(typo|hint|note|comment|TODO|FIXME)": -20,
    r"(style:|convention|one-?liner)": -20,
    r"(argument|absolute_import|attribute|assignment|mutable)": -15,
    r"(scope|True|False)": -10,
    r"(unused|useless|unnecessar|superfluous|duplicate|deprecated)": -10,
    r"(redundant|pointless|confusing|uninitialized|meaningless|dead)": -10,
    r": (drop|remove|delete|rip out)": -10,
    r": (inherit|rename|simplify|naming|inline)": -10,
    r"(correct doc|docstring|document .* method)": -20,
    r"(abstract|factor|extract|prepare|split|replace| import)": -20,
    r": add.*(function|method|implementation|example)": -10,
    r": (move|extract) .* (to|into|from|out of)": -20,
    r": implement ": -5,
    r": use .* implementation": -20,
    r": use .* instead of": -20,
    # code
    r"_": -10,
    r"__": -5,
    r"\(\)": -5,
    r"\S\S\S+\.\S\S\S\S+": -5,
    # dumb keywords
    r"\S+/\S+:": -10,
    r"\S+\.\S+:": -10,
    # python compatibility
    r"[Pp]y(|thon) ?[23]": -20,
    r"pycompat": -20,
    r"(coerce|convert|encode) .*to (byte|sys|)(s|str|string)": -20,
    # tests
    r"^test(|s|ing|runner|-\S+):": -20,
    r"^(f|hghave|run-tests):": -20,
    r"add.* tests?": -20,
    r"(buildbot|fuzz|mock|ratchet)": -10,
    # drop
    r"^i18n-": -50,
    r"^i18n:.*(hint|comment)": -50,
    r"perf:": -50,
    r"Added.*for changeset": -50,
    r"^_": -50,
}

cutoff = 10
commits = []

groupings = [
    (r"util|parsers|repo|ctx|context|revlog|filelog|alias|cmdutil", "core"),
    (r"revset|template|ui|dirstate|hook|i18n|transaction|wire|vfs", "core"),
    (r"dispatch|exchange|localrepo|streamclone|color|pager", "core"),
    (r"hgweb|paper|coal|gitweb|monoblue|spartan", "hgweb"),
    (r"pull|push|revert|resolve|annotate|bookmark|branch|clone", "commands"),
    (r"commands|commit|config|files|graft|import|log|merge|patch", "commands"),
    (r"phases|status|summary|amend|tag|help|verify", "commands"),
    (r"rebase|mq|convert|eol|histedit|largefiles", "extensions"),
    (r"shelve|unshelve", "extensions"),
]


def wikify(desc):
    desc = desc.replace("(issue", "(Bts:issue")
    desc = re.sub(r"\b([0-9a-f]{12})\b", r"Cset:\1", desc)
    # stop ParseError from being recognized as a (nonexistent) wiki page
    desc = re.sub(r" ([A-Z][a-z]+[A-Z][a-z]+)\b", r" !\1", desc)
    # prevent wiki markup of magic methods
    desc = re.sub(r"\b(\S*__\S*)\b", r"`\1`", desc)
    return desc


def main():
    desc = "example: %(prog)s 4.7.2 --stoprev 4.8rc0"
    ap = argparse.ArgumentParser(description=desc)
    ap.add_argument(
        "startrev",
        metavar="REV",
        type=str,
        help=(
            "Starting revision for the release notes. This revision "
            "won't be included, but later revisions will."
        ),
    )
    ap.add_argument(
        "--stoprev",
        metavar="REV",
        type=str,
        default="@",
        help=(
            "Stop revision for release notes. This revision will be included,"
            " but no later revisions will. This revision needs to be "
            "a descendant of startrev."
        ),
    )
    args = ap.parse_args()
    fromext = subprocess.check_output(
        [
            "hg",
            "--config",
            "extensions.releasenotes=",
            "releasenotes",
            "-r",
            "only(%s, %s)" % (args.stoprev, args.startrev),
        ]
    ).decode("utf-8")
    # Find all release notes from un-relnotes-flagged commits.
    for entry in sorted(
        subprocess.check_output(
            [
                "hg",
                "log",
                "-r",
                "only(%s, %s) - merge()" % (args.stoprev, args.startrev),
                "-T",
                r"{desc|firstline}\n",
            ]
        )
        .decode("utf-8")
        .splitlines()
    ):
        desc = entry.replace("`", "'")

        score = 0
        for rule, val in rules.items():
            if re.search(rule, desc):
                score += val

        if score >= cutoff:
            commits.append(wikify(desc))
    # Group unflagged notes.
    groups = {}
    bcs = []
    apis = []

    for d in commits:
        if "(BC)" in d:
            bcs.append(d)
        if "(API)" in d:
            apis.append(d)
        for rule, g in groupings:
            if re.match(rule, d):
                groups.setdefault(g, []).append(d)
                break
        else:
            groups.setdefault("unsorted", []).append(d)
    print(fromext)
    # print legacy release notes sections
    for g in sorted(groups):
        print("\n=== %s ===" % g)
        for d in sorted(groups[g]):
            print(" * %s" % d)

    if bcs:
        print("\n=== Behavior Changes ===\n")

    for d in sorted(bcs):
        print(" * %s" % d)

    if apis:
        print("\n=== Internal API Changes ===\n")

    for d in sorted(apis):
        print(" * %s" % d)


if __name__ == "__main__":
    main()